下面寫(xiě)一下，這兩個(gè)要是有no了，怎么恢復(fù)。。 

如果是slave_io_running no了，那么就我個(gè)人看有三種情況，一個(gè)是網(wǎng)絡(luò)有問(wèn)題，連接不上，像有一次我用虛擬機(jī)搭建replication，使用了nat的網(wǎng)絡(luò)結(jié)構(gòu)，就是死都連不上，第二個(gè)是有可能my.cnf有問(wèn)題，配置文件怎么寫(xiě)就不說(shuō)了，網(wǎng)上太多了，最后一個(gè)是授權(quán)的問(wèn)題，replication slave和file權(quán)限是必須的。如果不怕死就all咯。。 

一旦io為no了先看err日志，看看爆什么錯(cuò)，很可能是網(wǎng)絡(luò)，也有可能是包太大收不了，這個(gè)時(shí)候主備上改max_allowed_packet這個(gè)參數(shù)。 

如果是slave_sql_running no了，那么也有兩種可能，一種是slave機(jī)器上這個(gè)表中出現(xiàn)了其他的寫(xiě)操作，就是程序?qū)懥耍@個(gè)是會(huì)有問(wèn)題的，今天我想重現(xiàn)，但是有時(shí)候會(huì)有問(wèn)題，有時(shí)候就沒(méi)有問(wèn)題，現(xiàn)在還不是太明了，后面再更新，還有一種占絕大多數(shù)可能的是slave進(jìn)程重啟，事務(wù)回滾造成的，這也是mysql的一種自我保護(hù)的措施，像關(guān)鍵時(shí)候只讀一樣。 

這個(gè)時(shí)候想恢復(fù)的話，只要停掉slave，set GLOBAL SQL_SLAVE_SKIP_COUNTER=1;再開(kāi)一下slave就可以了，這個(gè)全局變量賦值為N的意思是： 

This statement skips the next N events from the master. This is useful for recovering from replication stops caused by a statement. 

This statement is valid only when the slave thread is not running. Otherwise, it produces an error. 

呵呵，講的比我清楚。 

MYSQL鏡像服務(wù)器因錯(cuò)誤停止的恢復(fù) 

下午主服務(wù)器，由于一些原因，導(dǎo)致死機(jī)，重啟后，發(fā)現(xiàn)從服務(wù)器的數(shù)據(jù)沒(méi)有跟上。 
配好MYSQL主從也才前幾天的事，沒(méi)多少經(jīng)驗(yàn)，第一次碰上這問(wèn)題，有點(diǎn)焦急。不過(guò)，自己試了下，還算解決了：） 

從服務(wù)器上 
Master_Log_File: mysqlhxmaster.000007 
Read_Master_Log_Pos: 84285377 

看一下主服務(wù)器：mysqlhxmaster.000007 | 84450528 | 
已經(jīng)過(guò)后很多了，確實(shí)沒(méi)跟上。 

show slave status\G 
Slave_IO_Running: Yes 
Slave_SQL_Running: No 

有問(wèn)題了，Slave_SQL_Running應(yīng)該是Yes才對(duì)。 


再往下看，有錯(cuò)誤的提示： 

Last_Errno: 1053 
Last_Error: Query partially completed on the master (error on master: 1053) and was aborted. There is a chance that your master is inconsistent at this point. If you are sure that your master is ok, run this query manually on the slave and then restart the slave with SET GLOBAL SQL_SLAVE_SKIP_COUNTER=1; START SLAVE; . Query: 'INSERT INTO hx_stat_record ......(一句SQL語(yǔ)句)' 

這里有說(shuō)明要怎么操作了：） 

先stop slave，然后執(zhí)行了一下提示的語(yǔ)句，再SET GLOBAL SQL_SLAVE_SKIP_COUNTER=1; START SLAVE; 

show slave status\G 

Slave_IO_Running: Yes 
Slave_SQL_Running: Yes 

OK了，從服務(wù)器也在幾分鐘內(nèi)把堆積的log處理完了，兩邊又同步了：) 

從MYSQL服務(wù)器Slave_IO_Running: No的解決2 

早晨機(jī)房意外斷電，導(dǎo)致了發(fā)現(xiàn)mysql從服務(wù)器同步異常。使用以前碰到的Slave_SQL_Running為No的解決辦法無(wú)效，仍然無(wú)法同步。 

查看一下?tīng)顟B(tài)show slave status 
Master_Log_File: mysqlmaster.000079 
Read_Master_Log_Pos: 183913228 
Relay_Log_File: hx-relay-bin.002934 
Relay_Log_Pos: 183913371 
Relay_Master_Log_File: mysqlmaster.000079 
Slave_IO_Running: No 
Slave_SQL_Running: Yes 

主服務(wù)器show master status\G 
File: mysqlmaster.000080 
Position: 13818288 
Binlog_Do_DB: 
Binlog_Ignore_DB: mysql,test 

mysql錯(cuò)誤日志： 
100512 9:13:17 [Note] Slave SQL thread initialized, starting replication in log 'mysqlmaster.000079' at position 183913228, relay log './hx-relay-bin.002934' position: 183913371 
100512 9:13:17 [Note] Slave I/O thread: connected to master '[email protected]:3306', replication started in log 'mysqlmaster.000079' at position 183913228 
100512 9:13:17 [ERROR] Error reading packet from server: Client requested master to start replication from impossible position ( server_errno=1236) 
100512 9:13:17 [ERROR] Got fatal error 1236: 'Client requested master to start replication from impossible position' from master when reading data from binary log 
100512 9:13:17 [Note] Slave I/O thread exiting, read up to log 'mysqlmaster.000079', position 183913228 

這次是Slave_IO_Running為No，從日志上來(lái)看，服務(wù)器讀mysqlmaster.000079這個(gè)Log的183913228這個(gè)位置時(shí)發(fā)生錯(cuò)誤，這個(gè)位置不存在，于是無(wú)法同步。 

查看一下這個(gè)Log的最后幾行： 
/*!40019 SET @@session.max_insert_delayed_threads=0*/; 
/*!50003 SET @OLD_COMPLETION_TYPE=@@COMPLETION_TYPE,COMPLETION_TYPE=0*/; 
# at 4 
#100511 9:35:15 server id 1 end_log_pos 98 Start: binlog v 4, server v 5.0.27-standard-log created 100511 9:35:15 
# Warning: this binlog was not closed properly. Most probably mysqld crashed writing it. 

嘗試從損壞之前的位置開(kāi)始 
SLAVE STOP; 
CHANGE MASTER TO MASTER_LOG_FILE='mysqlcncnmaster.000079', MASTER_LOG_POS=183913220; 
SLAVE START; 
無(wú)效！ 
只好從新的日志開(kāi)始 
SLAVE STOP; 
CHANGE MASTER TO MASTER_LOG_FILE='mysqlcncnmaster.000080', MASTER_LOG_POS=0; 
SLAVE START; 
此時(shí)Slave_IO_Running恢復(fù)為Yes，同步進(jìn)行了！觀察了會(huì)兒，沒(méi)有任何出錯(cuò)跡象，問(wèn)題解決。 

另外，出現(xiàn)Slave_IO_Running:NO還有一個(gè)原因是slave上沒(méi)有權(quán)限讀master上的數(shù)據(jù)。